warp.norm_huber#

warp.norm_huber(v: Any, delta: float) → float#

  • Kernel

  • Differentiable

Compute the Huber norm of a vector v with a given delta.

\[\begin{split}H(v) = \begin{cases} \frac{1}{2} \|v\|^2 & \text{if } \|v\| \leq \delta \\ \delta(\|v\| - \frac{1}{2}\delta) & \text{otherwise} \end{cases}\end{split}\]
../../_images/norm_huber.svg
Parameters:

  • v (Vector[Float, Any]) – The vector to compute the Huber norm of.

  • delta (float) – The threshold value, defaults to 1.0.

Returns:

The Huber norm of the vector.

Return type:

float